Home Forums Data Warehousing Analysis Services nable to build SQL Server OLAP Cube when there is no rows corresponding to foreign key in a table RE: nable to build SQL Server OLAP Cube when there is no rows corresponding to foreign key in a table

  • grasshopper26 (10/29/2012)


    Scott Murray-240410 (10/29/2012)


    grasshopper26 (10/28/2012)


    I am new to SQL Server OLAP Cubes. I am having the following issue like

    ex I have purchase order and invoice tables which are used in data source view. These two tables are related by purchase order ID which have one to many relationship with invoices.

    I am getting the following error for the purcahse orders which i dont have invoices

    Errors in the OLAP storage engine: The attribute key cannot be found when processing: Table: purchase order

    Can anyone throw some light on this to help me

    I would first check that your dimensions are fully processed with the most current data. If data has been added to the underlying source, but the dimensions have not been processed, that new data is not reflected in the dimension, so when the cube attempts to join the keys columns from fact to dimension, it is not available. If the dimension is up to date, then you will need to check the underlying source and find which invoices are missing. Generally, you should have a value for all keys or you can choose to how to process these error rows by choosing to ignore the missing values. I do not prefer this second option.

    How can i have the value for the missing invoices in invoice table should it be a dummy insert row ? if yes how is it going to effect the data in Excel UI for Summations?

    Not sure about dummy records, but if you have a PO then presumably there is one or more invoices for it, correct? Otherwise, the relationship is broken. Doesn't matter what you plan to view it with, it's broken, cube won't process.

    we do not use dummy records here. If there is a key in a fact table, that is because it looked up that key in the dimension table during fact ETL.

    If I get an error like you did, it means, as previous poster stated, that the dimension table may not have been processed. Do that first. If you still get the error, pluck out the query that processing is running and run it yourself to see the errors and do recon.

    If it isn't the dimension, double-check the relationship you have set up.